2003 Aerial Robotics
Atlanta
25 July 2003
Yesterday, six teams competed at the 2003 International Aerial Robotics Competition, at the McKenna MOUT site in Fort Benning, GA. The Georgia Tech Aerial Robotics (GTAR) team was one of three teams to successfully operate an autonomous air vehicle at the contest. Two of these teams accomplished “levels” of the contest mission. The University of Arizona accomplished “level 1” which involves the vehicle flying a 3 km course. The Georgia Tech team (having previously accomplished level 1) accomplished “level 2”, meaning that the GTAR team remains the leading team in the contest for the third year in a row.
Four attempts were made at the Level 2 mission, on three of those attempts, the GTAR research UAV system (which we refer to as the GTMax) automatically flew a search pattern over a simulated European village, looking at all of the walls of all of the buildings within. It automatically located a pre-specified sign on one of the buildings, identifying the correct “building of interest”. The vehicle then automatically flew a search pattern looking for openings into the building, and then gave the location of a valid opening to the contest judges along with a picture of that opening. In summary, the system was able to find a way into a specific building in a village identified only by a one meter wide sign on one wall. There were 15 buildings in the village.
The team consists of graduate and undergraduate students from the Georgia Tech schools of Aerospace Engineering and Electrical & Computer Engineering. Many of the participants are also part of our DARPA Software Enabled Control Program, with includes related UAV research that overlaps with these efforts. The team wishes to thank other sponsors of GTAR: Lockheed Martin, NovAtel, Guided Systems Technologies, and Inflight Robotics. More information on the contest can be found at: http://iarc.angel-strike.com/
More information on GTAR can be found at: http://controls.ae.gatech.edu/gtar/
Team that participated heavily in the contest this year:
- Henrik Christophersen, ECE undergraduate student, avionics hardware
- Jincheol Ha, AE graduate student, image processing
- Jeong Hur, AE research engineer, pilot and helicopter
- Suresh Kannan, AE graduate student, GN&C and avionics software
- Wayne Pickell, AE research engineer, avionics hardware
- Alison Proctor, AE graduate student, image processing and GN&C
- …and contributions from many others.
Unabridged list of pictures taken at the 2003 competition
Selected Media
Selected Movies
- f030723a1_symbolFindPractice.mpg
- symbol finding practice
- f030723a2_level2inPractice.mpg
- level 2 practice
- f030724a1_attempt1start.mpg
- starting 1st attempt at Level 2
- f030724b1_attempt2symbolFound.mpg
- sequence when symbol has been found
- f030724b2_attempt2correctBuilding.mpg
- identifying the correct building
- f030724c1_attempt3winningOpening.mpg
- finding the winning open portal
- f030724c2_attempt4searching.mpg
- start searching
- f030724c3_attempt4autolandFinish.mpg
- auto land at end of last attempt
- obgtar2003attempt2_5x.avi
- avi made from still images from onboard camera
- obgtar2003attempt3_5x.avi
- avi made from still images from onboard camera
- l030724a1_level3launcherTest.mpg
- test of the launcher mechanism
Selected Pictures
Hardware in the loop setup with helicopter performing entire mission in simulation. The camera of the helicopter is pointed at a computer generated image of what the camera would see.
A view of the main groundstation during the mission.
A view of the image processing ground station during the mission. All image processing is performed onboard the helicopter, the ground station is only used to display results for the judges.
Output of image processor after it identified the symbol during the competition.
The GTMax during the mission.
The Team.
Setting up the GTMax the night before in darkness and rain.
Almost like a scene from SGI‘s OpenGL Performer town
Just before contest begins
Testing Henrik’s launch contraption